Suppose a puzzle says: "Two numbers add to 5. The same two numbers add to 8." Whatever the numbers are, x+y is one quantity, and one quantity cannot be both 5 and 8. Subtracting the equations makes this plain: (x+y)−(x+y)=8−5 becomes 0=3. The variables have vanished, and what remains is false whatever x and y are. That false, variable-free statement is a contradiction, and it means the puzzle has no solution.
The logic behind it: if some pair of numbers did satisfy both equations, every correct step would keep producing true statements about that pair. A correct step that produces a false statement shows that no such pair exists. Proof by contradiction runs on the same logic: assume something, reason correctly, reach something impossible, and conclude that the assumption was wrong.
Not every variable-free result is a contradiction. 0=0 is always true; it means one equation repeated another, and the system may have infinitely many solutions. An equation like 2x=6 is a conditional equation, true for x=3 only, so it is not a contradiction either. A contradiction is false for every value, and you can trust one only after checking that the steps leading to it were correct.
Think of it like this. A contradiction works like a broken alibi: if a suspect was in Paris at noon and also in Rome at noon, the story is impossible, whoever the suspect is. Reaching 0=3 is that impossible alibi for the variables. As with an alibi, check the evidence first, since a misread clock can create a conflict that is not real.
Core idea. A contradiction is a claim no values can ever satisfy, like 0=3.